Inceptionv3, compute batch size from --gpu option instead of hard-coded. (#246)

examples/Inception/inception-bn.py

@@ -192,4 +192,5 @@ if __name__ == '__main__':
         config.session_init = SaverRestore(args.load)
     if args.gpu:
         config.nr_tower = len(args.gpu.split(','))
+        assert config.nr_tower == NR_GPU
     SyncMultiGPUTrainer(config).train()

examples/Inception/inceptionv3.py

@@ -22,13 +22,11 @@ This config follows the official inceptionv3 setup
 (https://github.com/tensorflow/models/tree/master/inception/inception)
 with much much fewer lines of code.
 It reaches 74% single-crop validation accuracy, similar to the official code.
-
-The hyperparameters here are for 8 GPUs, so the effective batch size is 8*64 = 512.
 """
 
 TOTAL_BATCH_SIZE = 512
-NR_GPU = 8
-BATCH_SIZE = TOTAL_BATCH_SIZE // NR_GPU
+NR_GPU = None
+BATCH_SIZE = None
 INPUT_SHAPE = 299
 
 
@@ -285,19 +283,19 @@ def get_config():
 
 if __name__ == '__main__':
     parser = argparse.ArgumentParser()
-    parser.add_argument('--gpu', help='comma separated list of GPU(s) to use.')
+    parser.add_argument('--gpu', help='comma separated list of GPU(s) to use.', required=True)
     parser.add_argument('--data', help='ILSVRC dataset dir')
     parser.add_argument('--load', help='load model')
     args = parser.parse_args()
 
     logger.auto_set_dir()
 
-    if args.gpu:
-        os.environ['CUDA_VISIBLE_DEVICES'] = args.gpu
+    os.environ['CUDA_VISIBLE_DEVICES'] = args.gpu
+    NR_GPU = len(args.gpu.split(','))
+    BATCH_SIZE = TOTAL_BATCH_SIZE // NR_GPU
 
     config = get_config()
     if args.load:
         config.session_init = SaverRestore(args.load)
-    if args.gpu:
-        config.nr_tower = len(args.gpu.split(','))
+    config.nr_tower = NR_GPU
     SyncMultiGPUTrainer(config).train()

examples/ResNet/imagenet-resnet.py

@@ -235,7 +235,7 @@ def eval_on_ILSVRC12(model_file, data_dir):
 
 if __name__ == '__main__':
     parser = argparse.ArgumentParser()
-    parser.add_argument('--gpu', help='comma separated list of GPU(s) to use.')
+    parser.add_argument('--gpu', help='comma separated list of GPU(s) to use.', required=True)
     parser.add_argument('--data', help='ILSVRC dataset dir')
     parser.add_argument('--load', help='load model')
     parser.add_argument('--fake', help='use fakedata to test or benchmark this model', action='store_true')
@@ -247,14 +247,13 @@ if __name__ == '__main__':
     args = parser.parse_args()
 
     DEPTH = args.depth
-    if args.gpu:
-        os.environ['CUDA_VISIBLE_DEVICES'] = args.gpu
+    os.environ['CUDA_VISIBLE_DEVICES'] = args.gpu
+
     if args.eval:
         BATCH_SIZE = 128    # something that can run on one gpu
         eval_on_ILSVRC12(args.load, args.data)
         sys.exit()
 
-    assert args.gpu is not None, "Need to specify a list of gpu for training!"
     NR_GPU = len(args.gpu.split(','))
     BATCH_SIZE = TOTAL_BATCH_SIZE // NR_GPU
 

tensorpack/dataflow/image.py

@@ -9,6 +9,7 @@ from .base import RNGDataFlow
 from .common import MapDataComponent, MapData
 from .imgaug import AugmentorList
 from ..utils import logger
+from ..utils.argtools import shape2d
 
 __all__ = ['ImageFromFile', 'AugmentImageComponent', 'AugmentImageComponents']
 
@@ -20,13 +21,13 @@ class ImageFromFile(RNGDataFlow):
         Args:
             files (list): list of file paths.
             channel (int): 1 or 3. Will convert grayscale to RGB images if channel==3.
-            resize (tuple): (h, w). If given, resize the image.
+            resize (tuple): int or (h, w) tuple. If given, resize the image.
         """
         assert len(files), "No image files given to ImageFromFile!"
         self.files = files
         self.channel = int(channel)
         self.imread_mode = cv2.IMREAD_GRAYSCALE if self.channel == 1 else cv2.IMREAD_COLOR
-        self.resize = resize
+        self.resize = shape2d(resize)
         self.shuffle = shuffle
 
     def size(self):
@@ -40,7 +41,7 @@ class ImageFromFile(RNGDataFlow):
             if self.channel == 3:
                 im = im[:, :, ::-1]
             if self.resize is not None:
-                im = cv2.resize(im, self.resize[::-1])
+                im = cv2.resize(im, tuple(self.resize[::-1]))
             if self.channel == 1:
                 im = im[:, :, np.newaxis]
             yield [im]